annotate src/fftw-3.3.3/dft/scalar/codelets/n1_3.c @ 169:223a55898ab9 tip default

Add null config files
author Chris Cannam <cannam@all-day-breakfast.com>
date Mon, 02 Mar 2020 14:03:47 +0000
parents 89f5e221ed7b
children
rev   line source
cannam@95 1 /*
cannam@95 2 * Copyright (c) 2003, 2007-11 Matteo Frigo
cannam@95 3 * Copyright (c) 2003, 2007-11 Massachusetts Institute of Technology
cannam@95 4 *
cannam@95 5 * This program is free software; you can redistribute it and/or modify
cannam@95 6 * it under the terms of the GNU General Public License as published by
cannam@95 7 * the Free Software Foundation; either version 2 of the License, or
cannam@95 8 * (at your option) any later version.
cannam@95 9 *
cannam@95 10 * This program is distributed in the hope that it will be useful,
cannam@95 11 * but WITHOUT ANY WARRANTY; without even the implied warranty of
cannam@95 12 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
cannam@95 13 * GNU General Public License for more details.
cannam@95 14 *
cannam@95 15 * You should have received a copy of the GNU General Public License
cannam@95 16 * along with this program; if not, write to the Free Software
cannam@95 17 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
cannam@95 18 *
cannam@95 19 */
cannam@95 20
cannam@95 21 /* This file was automatically generated --- DO NOT EDIT */
cannam@95 22 /* Generated on Sun Nov 25 07:35:42 EST 2012 */
cannam@95 23
cannam@95 24 #include "codelet-dft.h"
cannam@95 25
cannam@95 26 #ifdef HAVE_FMA
cannam@95 27
cannam@95 28 /* Generated by: ../../../genfft/gen_notw.native -fma -reorder-insns -schedule-for-pipeline -compact -variables 4 -pipeline-latency 4 -n 3 -name n1_3 -include n.h */
cannam@95 29
cannam@95 30 /*
cannam@95 31 * This function contains 12 FP additions, 6 FP multiplications,
cannam@95 32 * (or, 6 additions, 0 multiplications, 6 fused multiply/add),
cannam@95 33 * 15 stack variables, 2 constants, and 12 memory accesses
cannam@95 34 */
cannam@95 35 #include "n.h"
cannam@95 36
cannam@95 37 static void n1_3(const R *ri, const R *ii, R *ro, R *io, stride is, stride os, INT v, INT ivs, INT ovs)
cannam@95 38 {
cannam@95 39 DK(KP866025403, +0.866025403784438646763723170752936183471402627);
cannam@95 40 DK(KP500000000, +0.500000000000000000000000000000000000000000000);
cannam@95 41 {
cannam@95 42 INT i;
cannam@95 43 for (i = v; i > 0; i = i - 1, ri = ri + ivs, ii = ii + ivs, ro = ro + ovs, io = io + ovs, MAKE_VOLATILE_STRIDE(12, is), MAKE_VOLATILE_STRIDE(12, os)) {
cannam@95 44 E T1, T9, T2, T3, T6, T7;
cannam@95 45 T1 = ri[0];
cannam@95 46 T9 = ii[0];
cannam@95 47 T2 = ri[WS(is, 1)];
cannam@95 48 T3 = ri[WS(is, 2)];
cannam@95 49 T6 = ii[WS(is, 1)];
cannam@95 50 T7 = ii[WS(is, 2)];
cannam@95 51 {
cannam@95 52 E T4, Tc, T8, Ta, T5, Tb;
cannam@95 53 T4 = T2 + T3;
cannam@95 54 Tc = T3 - T2;
cannam@95 55 T8 = T6 - T7;
cannam@95 56 Ta = T6 + T7;
cannam@95 57 T5 = FNMS(KP500000000, T4, T1);
cannam@95 58 ro[0] = T1 + T4;
cannam@95 59 Tb = FNMS(KP500000000, Ta, T9);
cannam@95 60 io[0] = T9 + Ta;
cannam@95 61 ro[WS(os, 1)] = FMA(KP866025403, T8, T5);
cannam@95 62 ro[WS(os, 2)] = FNMS(KP866025403, T8, T5);
cannam@95 63 io[WS(os, 2)] = FNMS(KP866025403, Tc, Tb);
cannam@95 64 io[WS(os, 1)] = FMA(KP866025403, Tc, Tb);
cannam@95 65 }
cannam@95 66 }
cannam@95 67 }
cannam@95 68 }
cannam@95 69
cannam@95 70 static const kdft_desc desc = { 3, "n1_3", {6, 0, 6, 0}, &GENUS, 0, 0, 0, 0 };
cannam@95 71
cannam@95 72 void X(codelet_n1_3) (planner *p) {
cannam@95 73 X(kdft_register) (p, n1_3, &desc);
cannam@95 74 }
cannam@95 75
cannam@95 76 #else /* HAVE_FMA */
cannam@95 77
cannam@95 78 /* Generated by: ../../../genfft/gen_notw.native -compact -variables 4 -pipeline-latency 4 -n 3 -name n1_3 -include n.h */
cannam@95 79
cannam@95 80 /*
cannam@95 81 * This function contains 12 FP additions, 4 FP multiplications,
cannam@95 82 * (or, 10 additions, 2 multiplications, 2 fused multiply/add),
cannam@95 83 * 15 stack variables, 2 constants, and 12 memory accesses
cannam@95 84 */
cannam@95 85 #include "n.h"
cannam@95 86
cannam@95 87 static void n1_3(const R *ri, const R *ii, R *ro, R *io, stride is, stride os, INT v, INT ivs, INT ovs)
cannam@95 88 {
cannam@95 89 DK(KP500000000, +0.500000000000000000000000000000000000000000000);
cannam@95 90 DK(KP866025403, +0.866025403784438646763723170752936183471402627);
cannam@95 91 {
cannam@95 92 INT i;
cannam@95 93 for (i = v; i > 0; i = i - 1, ri = ri + ivs, ii = ii + ivs, ro = ro + ovs, io = io + ovs, MAKE_VOLATILE_STRIDE(12, is), MAKE_VOLATILE_STRIDE(12, os)) {
cannam@95 94 E T1, Ta, T4, T9, T8, Tb, T5, Tc;
cannam@95 95 T1 = ri[0];
cannam@95 96 Ta = ii[0];
cannam@95 97 {
cannam@95 98 E T2, T3, T6, T7;
cannam@95 99 T2 = ri[WS(is, 1)];
cannam@95 100 T3 = ri[WS(is, 2)];
cannam@95 101 T4 = T2 + T3;
cannam@95 102 T9 = KP866025403 * (T3 - T2);
cannam@95 103 T6 = ii[WS(is, 1)];
cannam@95 104 T7 = ii[WS(is, 2)];
cannam@95 105 T8 = KP866025403 * (T6 - T7);
cannam@95 106 Tb = T6 + T7;
cannam@95 107 }
cannam@95 108 ro[0] = T1 + T4;
cannam@95 109 io[0] = Ta + Tb;
cannam@95 110 T5 = FNMS(KP500000000, T4, T1);
cannam@95 111 ro[WS(os, 2)] = T5 - T8;
cannam@95 112 ro[WS(os, 1)] = T5 + T8;
cannam@95 113 Tc = FNMS(KP500000000, Tb, Ta);
cannam@95 114 io[WS(os, 1)] = T9 + Tc;
cannam@95 115 io[WS(os, 2)] = Tc - T9;
cannam@95 116 }
cannam@95 117 }
cannam@95 118 }
cannam@95 119
cannam@95 120 static const kdft_desc desc = { 3, "n1_3", {10, 2, 2, 0}, &GENUS, 0, 0, 0, 0 };
cannam@95 121
cannam@95 122 void X(codelet_n1_3) (planner *p) {
cannam@95 123 X(kdft_register) (p, n1_3, &desc);
cannam@95 124 }
cannam@95 125
cannam@95 126 #endif /* HAVE_FMA */